rwd-cohort-analysis

Pass

Audited by Gen Agent Trust Hub on Jul 9, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns detected. The skill strictly provides boilerplate code for clinical cohort identification, medication adherence calculation, and survival analysis using established libraries.
  • [SAFE]: Data handling is performed locally on user-provided data structures. There are no network requests, external data fetches, or interactions with untrusted remote sources.
  • [SAFE]: All referenced dependencies are industry-standard packages for data science and biostatistics (e.g., pandas, scikit-learn, lifelines, MatchIt, survival).
  • [SAFE]: No obfuscation, prompt injection, or privilege escalation patterns were identified in the instructions or code snippets.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 9, 2026, 06:00 PM
Security Audit — agent-trust-hub — rwd-cohort-analysis